As shown in the previous chapters, the model's predictions can vary depending on the value of **k** (the number of neighbors). When building a k-NN model, it's important to choose the k value that gives the **best performance**.

A common approach is to use **cross-validation** to evaluate model performance. You can run a loop and calculate cross-validation scores **for a range of k values**, then select the one with the highest score. This is the most widely used method.

To perform this, `sklearn` offers a convenient tool: the `GridSearchCV` class.

The `param_grid` parameter takes a **dictionary** where the keys are parameter names and the values are lists of options to try. For example, to test values from `1` to `99` for `n_neighbors`, you can write:
```python
param_grid = {'n_neighbors': range(1, 100)}
```
Calling the `.fit(X, y)` method on the `GridSearchCV` object will search through the parameter grid to find the best parameters and then **re-train the model on the entire dataset** using those best parameters.

You can access the **best score** using the `.best_score_` attribute and make predictions with the optimized model using the `.predict()` method. Similarly, you can retrieve the **best model** itself using the `.best_estimator_` attribute.

import unittest
import importlib


def _dynamic_test(test_case, condition, success_message, failure_message):
    if condition:
        test_case._testMethodName = success_message
        test_case.assertTrue(True, success_message)
    else:
        test_case._testMethodName = failure_message
        test_case.fail(failure_message)


class TestUserCode(unittest.TestCase):

    def test_param_grid_is_declared(self):
        import user_code

        variable = 'param_grid'
        _dynamic_test(
            self,
            hasattr(user_code, variable),
            f"The `{variable}` variable is declared.",
            f"Expected `{variable}` to be declared."
        )

    def test_para_grid_is_dict(self):
        import user_code

        try:
            condition = isinstance(user_code.param_grid, dict)
            failure_message = f"Expected `param_grid` to be a `dict`, but got `{type(user_code.param_grid).__name__}`."
        except AttributeError:
            condition = False
            failure_message = "The `param_grid` variable is not declared."

        _dynamic_test(
            self,
            condition,
            "`param_grid` is a `dict`.",
            failure_message
        )

    def test_param_grid_is_correct(self):
        import user_code
        expected_value = {'n_neighbors': [3, 9, 18, 27]}

        variable = 'param_grid'
        actual_value = getattr(user_code, variable, None)
        if actual_value is None:
            condition = False
            failure_message = f"The `{variable}` variable is not declared."
        elif isinstance(actual_value, dict):
            condition = actual_value == expected_value
            failure_message = f"Expected `{variable}` to contain `{expected_value}`, but got `{actual_value}`."
        else:
            condition = False
            failure_message = f"`{variable}` is not a `dict`."

        _dynamic_test(
            self,
            condition,
            f"`{variable}` contains the correct values.",
            failure_message
        )

    def test_grid_search_is_declared(self):
        import user_code

        variable = 'grid_search'
        _dynamic_test(
            self,
            hasattr(user_code, variable),
            f"The `{variable}` variable is declared.",
            f"Expected `{variable}` to be declared."
        )

    def test_grid_search_is_correct(self):
        import user_code
        from sklearn.model_selection import GridSearchCV
        from sklearn.neighbors import KNeighborsClassifier

        knn = KNeighborsClassifier()
        param_grid = {'n_neighbors': [3, 9, 18, 27]}

        variable = 'grid_search'
        actual_value = getattr(user_code, variable, None)
        if actual_value is None:
            condition = False
            failure_message = f"The `{variable}` variable is not declared."
        elif isinstance(actual_value, GridSearchCV):
            condition = isinstance(actual_value.estimator, KNeighborsClassifier) and actual_value.param_grid == param_grid and actual_value.cv == 4
            failure_message = f"Expected `{variable}` to be a `GridSearchCV` with `knn` as the model, `4` folds, and `param_grid` as the grid of parameters, but got `estimator={actual_value.estimator}`, `param_grid={actual_value.param_grid}`, and `cv={actual_value.cv}`."
        else:
            condition = False
            failure_message = f"`{variable}` is not a `GridSearchCV`."

        _dynamic_test(
            self,
            condition,
            "`grid_search` is a `GridSearchCV` with `estimator=knn`, `param_grid=param_grid`, and `cv=4`.",
            failure_message
        )

    def test_best_model_is_declared(self):
        import user_code

        variable = 'best_model'
        _dynamic_test(
            self,
            hasattr(user_code, variable),
            f"The `{variable}` variable is declared.",
            f"Expected `{variable}` to be declared."
        )

    def test_best_model_is_correct(self):
        import user_code
        from sklearn.neighbors import KNeighborsClassifier

        variable = 'best_model'
        actual_value = getattr(user_code, variable, None)
        if actual_value is None:
            condition = False
            failure_message = f"The `{variable}` variable is not declared."
        elif isinstance(actual_value, KNeighborsClassifier):
            condition = actual_value.n_neighbors == 27
            failure_message = f"Expected `{variable}` to be a `KNeighborsClassifier` with `27` neighbors, but got `n_neighbors={actual_value.n_neighbors}`."
        else:
            condition = False
            failure_message = f"`{variable}` is not a `KNeighborsClassifier`."

        _dynamic_test(
            self,
            condition,
            "`best_model` is a `KNeighborsClassifier` with `27` neighbors.",
            failure_message
        )

    def test_best_score_is_declared(self):
        import user_code

        variable = 'best_score'
        _dynamic_test(
            self,
            hasattr(user_code, variable),
            f"The `{variable}` variable is declared.",
            f"Expected `{variable}` to be declared."
        )

    def test_best_score_is_correct(self):
        import user_code
        expected_value = 0.8039505464098539

        variable = 'best_score'
        actual_value = getattr(user_code, variable, None)
        if actual_value is None:
            condition = False
            failure_message = f"The `{variable}` variable is not declared."
        elif isinstance(actual_value, float):
            condition = actual_value == expected_value
            failure_message = f"Expected `{variable}` to contain `{expected_value}`, but got `{actual_value}`."
        else:
            condition = False
            failure_message = f"`{variable}` is not a `float`."

        _dynamic_test(
            self,
            condition,
            f"`{variable}` is correct.",
            failure_message
        )


if __name__ == '__main__':
    unittest.main()

test_main.py

Master the core classification algorithms that power modern machine learning. Explore how models like k-NN, logistic regression, decision trees, and random forests make predictions, evaluate their accuracy, and understand when to use each. Build the skills to compare models and choose the best one for your data.

Discover how the k-nearest neighbors algorithm makes predictions based on similarity. Learn to handle multiple features, tune parameters, and apply cross-validation to improve accuracy.

Understand how logistic regression models probabilities and classifies outcomes. Practice implementing it, interpreting decision boundaries, and applying regularization to prevent overfitting.

Learn how decision trees split data into meaningful groups based on feature values. Explore how parameters like tree depth and minimum samples per leaf affect model performance and generalization.

Explore how random forests combine multiple decision trees to improve accuracy and robustness. Understand the role of randomness and apply this ensemble method to real-world data.

Evaluate models using metrics such as accuracy, precision, recall, and F1-score. Learn to interpret confusion matrices and compare multiple classifiers to identify the best-performing model.

Challenge: Choosing the Best K Value

解答